APHERESIS COLUMN FOR TREATING RHEUMATOID ARTHRITIS

Abstract

The present invention relates to an apheresis column loaded with a solid support comprising a composition comprising at least one peptide selected from the group consisting of: —the αI7I-I85cit peptide of amino acid sequence VDIDIKIX.sub.1SCX.sub.2GSCS (SEQ ID NO: 8) wherein X.sub.1 and X.sub.2 each represent a citmllyl residue, —the α62I-635Cit peptide of amino acid sequence X1GHAKSX2PVX3GIHTS (SEQ ID NO: 12) wherein X1, X2 and X3 each represent a citmllyl residue—the P60-74cit-NH2 peptide of amino acid sequence X1PAPPPISGGGYX2AX3 (SEQ ID NO: 15) wherein X1 and X2 each represent a citmllyl residue and X3 represents a citmllyl derivative with a carboxamide group and—the peptide, referred to as the Ac-a36-50crt peptide, having the amino acid sequence GPX1VVEX2HQSACKDS (SEQ ID NO: 6) wherein the residue G at the N-terminal is acetylated and wherein X1 and X2 each represent a citmllyl residue and/or the a36-50cit peptide of amino acid sequence GPX1VVEX2HQSACKDS (SEQ ID NO: 5) wherein X.sub.1 and X.sub.2 each represent a citmllyl residue.

Claims

1. An apheresis column loaded with a solid support comprising a composition which comprises at least one peptide selected from the group consisting of: a peptide, α171-185.sub.Cit, comprising amino acid sequence VDIDIKIX.sub.1SCX.sub.2GSCS (SEQ ID NO: 8) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue, a peptide, α621-635.sub.Cit, comprising amino acid sequence X.sub.1GHAKSX.sub.2PVX.sub.3GIHTS (SEQ ID NO: 12) wherein X.sub.1, X.sub.2 and X.sub.3 each represent a citrullyl residue, a peptide, β60-74.sub.Cit-NH2, comprising amino acid sequence X.sub.1PAPPPISGGGYX.sub.2AX.sub.3 (SEQ ID NO: 15) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue and X.sub.3 represents a citrullyl derivative with a carboxamide group in place of the carboxyl group, and a peptide, .sub.Ac-α36-50.sub.Cit, comprising amino acid sequence GPX.sub.1VVEX.sub.2HQSACKDS (SEQ ID NO: 6) wherein the N-terminal G residue is acetylated and wherein X.sub.1 and X.sub.2 each represent a citrullyl residue, and/or a peptide, α36-50.sub.Cit, comprising amino acid sequence GPX.sub.1VVEX.sub.2HQSACKDS (SEQ ID NO: 5) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue; wherein the peptide or peptides are immobilised directly or indirectly on the solid support.

2. The apheresis column according to claim 1 wherein it comprises at least 2 peptides selected from the group consisting of α36-50.sub.Cit, .sub.Ac-α36-50.sub.Cit, α171-185.sub.Cit, α621-635.sub.Cit and β60-74.sub.Cit-NH2.

3. The apheresis column according to claim 1 comprising the α171-185.sub.Cit, α621-635.sub.Cit, β60-74.sub.Cit-NH2 peptides and the α36-50.sub.Cit and/or the .sub.Ac-α36-50.sub.Cit peptide or peptides.

4. The apheresis column according to claim 1 further comprising α501-515.sub.Cit peptide comprising amino acid sequence SGIGTLDGFX.sub.1HX.sub.2HPD (SEQ ID NO: 10) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue.

5. The apheresis column according to claim 1 comprising: the peptide or the peptides selected from the group consisting of α171-185.sub.Cit, α621-635.sub.Cit, and β60-74.sub.Cit-NH2, and optionally the α501-515.sub.Cit peptide, immobilised directly or indirectly on the solid support by their N-terminal end and/or the .sub.Ac-α36-50.sub.Cit and/or α36-50.sub.Cit peptide or peptides immobilised directly or indirectly on the solid support by their C-terminal end.

6. The apheresis column according to claim 1 that comprises the peptide or the peptides immobilised indirectly on the solid support by means of one or more intermediate compounds.

7. A composition comprising at least two peptides selected from the group consisting of: a peptide, α171-185.sub.Cit, comprising amino acid sequence VDIDIKIX.sub.1SCX.sub.2GSCS (SEQ ID NO: 8) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue, a peptide, α621-635.sub.Cit, comprising amino acid sequence X.sub.1GHAKSX.sub.2PVX.sub.3GIHTS (SEQ ID NO: 12) wherein X.sub.1, X.sub.2 and X.sub.3 each represent a citrullyl residue, a peptide, β60-74.sub.Cit-NH2, comprising amino acid sequence X.sub.1PAPPPISGGGYX.sub.2AX.sub.3 (SEQ ID NO: 15) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue and X.sub.3 represents a citrullyl derivative with a carboxamide group in place of the carboxyl group, and a peptide, .sub.Ac-α36-50.sub.Cit, comprising amino acid sequence GPX.sub.1VVEX.sub.2HQSACKDS (SEQ ID NO: 6) wherein the N-terminal G residue is acetylated and wherein X.sub.1 and X.sub.2 each represent a citrullyl residue and/or a peptide, α36-50.sub.Cit, comprising amino acid sequence GPX.sub.1VVEX.sub.2HQSACKDS (SEQ ID NO: 5) wherein X.sub.1 and X.sub.2 each represent a citrullyl residue.

8. Manufacturing an apheresis column comprising linking the composition as defined in claim 1 to a solid support.

9. A method for the treatment or the prevention of an autoimmune disease; comprising the administration of anti-citrullinated protein autoantibodies according to claim 1, wherein the peptide or the peptides of the composition are immobilised on a solid support.

10. The method according to claim 9, wherein the autoimmune disease is selected from the group of Sjögren syndrome, juvenile idiopathic arthritis, and rheumatoid arthritis.

11. The method according to claim 9 wherein, the treated subject has a high titer of anti-citrullinated protein autoantibodies (ACPA).

12. The method according to claim 10, wherein the autoimmune disease is rheumatoid arthritis.

13. The apheresis column according to claim 2, comprising the α171-185.sub.Cit, α621-635.sub.Cit, and β60-74.sub.Cit-NH2 peptides and the α36-50.sub.Cit and/or .sub.Ac-α36-50.sub.Cit peptide or peptides.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0096] FIG. 1 Distribution of the AhFibA titers determined on serum samples from 202 AhFibA-positive rheumatoid arthritis patients. The threshold values of Optical Density (the corrected OD defines the Titer) separating the quartiles are noted.

[0097] FIG. 2 Percentage of serums reactive in ELISA with respect to the 4 β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides, among 202 serums from AhFibA-positive rheumatoid arthritis patients.

[0098] FIG. 3 Distribution of 202 AhFibA-positive serums from rheumatoid arthritis patients according to their reactivity profile with respect to the 4 β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides.

[0099] FIG. 4 Correlations between the titers of the antibodies against the 4 β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides, for the 202 serums tested. At each intersection the Spearman coefficient and (p) of the correlation are noted.

[0100] FIG. 5 Two types of chromatography were used to simultaneously purify the anti-β60-74.sub.Cit-NH2, anti-.sub.Ac-α36-50.sub.Cit, anti-α621-635.sub.Cit and anti-α171-185.sub.Cit antibodies: either 4 mono-peptide columns connected in series (A), or a single column containing a balanced mixture of the 4 mono-peptide matrices (B).

[0101] FIG. 6 Chromatographs made on pools of serums from RA patients, either on the 4 serially mounted columns, loaded with the β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides respectively (left histograms), or on a column containing the mixture of the 4 peptides (right histograms). The pool 97 consisted of 97 serum samples of the same volume from patients with increasing ACPA titers representative of those found in a RA patient population; the Pool 27 consisted of 27 serum samples of the same volume from patients with high ACPA titers (OD≥1.5). The pools (starting serum: SD—black bar), the non-retained fractions (FNR—white bar) and the eluates (grey bar), were analysed in AhFibA ELISA (n≥24), the results are expressed as corrected OD.

[0102] FIG. 7 Method for calculating the purification rate of a serum after chromatography. The non-retained fraction is analysed in AhFibA ELISA and the OD obtained is compared to those resulting from the progressive dilution of the starting serum (reference curve). The ratio of dilutions allows to determine the purification rate. In this example, 1/50 vs 1/400: 8 times less antibodies in the FNR, i.e. 12.5% of the antibodies present in the starting serum. 87.5% of the ACPA were therefore purified.

[0103] FIG. 8 Calculation of the ACPA purification rate of a serum (n°41) after chromatography on the 4 columns loaded respectively with the β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit, α171-185.sub.Cit peptides mounted in series. The non-retained fraction was analysed in AhFibA ELISA and the OD obtained compared to those resulting from the dilution of the starting serum (reference curve). The ratio of 1/50 to 1/1718 dilutions allows to determine the purification rate: 98% of the ACPA have been purified.

[0104] FIG. 9 Chromatography made from the Pool 97, the Pool 27 and 10 individual serums, from RA patients on the 4 columns loaded with β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides respectively and mounted in series. For each pool and individual serum, the starting serums (SD—black bars) and the non-retained fractions (FNR—white bars) were analysed in AhFibA ELISA (n≥24): their titers are expressed as corrected OD. The purification rates (%) were calculated for all the chromatographed samples (grey bars).

[0105] FIG. 10 Correlation between the titer in ACPA (AhFibA ELISA) of the starting serum (before chromatography) and the purification rate obtained after chromatography on the 4 columns in series, loaded respectively with the β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides. The correlation is highly significant (p=0.014): the higher the ACPA titer of the serums, the more efficient the purification.

EXAMPLES

[0106] Materials and Methods

[0107] Serum Samples

[0108] Serum samples were obtained from 202 rheumatoid arthritis patients positive for the “AhFibA” ELISA test, which allows to detect and titer the ACPA using the citrullinated human fibrinogen as an immunosorbent.

[0109] Synthetic Peptides

[0110] Peptides corresponding to residues 36 to 50, 171 to 185 and 621 to 635 of the α-chain of the human fibrinogen (NP_068657—Isoform 2) and residues 60 to 74 of its β-chain (AAA18024) were synthesised.

[0111] The peptides were synthesised in their citrullinated form (by systematic substitution of all the arginyl residues by a citrullyl residue) or in their native non-citrullinated form (arginyl residues).

[0112] The β60-74 and β60-74.sub.Cit peptides were synthesised with a carboxamide in place of the terminal carboxyl group to give the β60-74.sub.NH2 and β60-74.sub.Cit-NH2 peptides.

[0113] The α36-50, and α36-50.sub.Cit peptides were synthesised with an acetyl on the N-terminal side to obtain the .sub.Ac-α36-50.sub.and Ac-α36-50.sub.Cit peptides.

[0114] The peptides were biotinylated at the C-terminal of an amino hexanoyl spacer for the .sub.Ac-α36-50Cit peptide and at the N-terminal of the same spacer, for the α171-185.sub.Cit, α621-635.sub.Cit and β60-74.sub.Cit-NH2 peptides.

[0115] Tests ELISA

[0116] The antigens or immunosorbents (citrullinated fibrinogen or citrullinated and non-citrullinated peptides), solubilised in PBS (1.5 mM KH.sub.2PO4 SIGMA 795488; 7 mM K.sub.2HPO.sub.4 SIGMA P3786; 0.15M NaCl SIGMA 31434; pH 7.2) at concentrations of 5 and 10 μg/mL, are adsorbed in the microtiter plate wells (MAXISORP NUNC 2023-09) by incubation overnight at 4° C., or used in avidin-biotin system: avidin 5 μg/mL in PBS incubated overnight at 4° C. and then, after washing, biotinylated peptide at 10 μg/mL in PBS, incubated 1 h at 4° C. After blocking in PBS 2% BSA (Sigma A3059) for 1 hour at 4° C. and washing in PBS 0.1% Tween®20 (SIGMA P1379), the samples to be tested are deposited diluted 1:50 (1:100 in the Avidin-Biotin system) or at equivalent corrected dilutions, in PBS 2% BSA 2MNaCl. After washing, the secondary antibody (goat anti-human IgG Fc fragment coupled to horseradish peroxidase; Southern Biotech 2040-05) is incubated for 1 hour at 4° C. After washing, the revelation is performed with a solution of Ortho-Phenylene Diamine dihydrochloride (Sigma P2788), 0.03% H.sub.2O.sub.2 (Sigma H1009), in citrate/H.sub.3PO.sub.4 buffer pH5 (0.05 M citric acid Sigma C0759; 0.1 M Na2HPO4 Prolabo 28026292) for 5 minutes at room temperature, and then stopped with a solution of H.sub.2SO.sub.4 6N (Sigma 07208). The optical density (OD) is read at 492 nm with a Multiskan Fc plate reader (ThermoScientific).

[0117] Conditions for Carrying ELISA Tests

TABLE-US-00002 TABLE 2 Concentration Secondary of the Sample antibody immunosorbent dilution dilution AhFibA  5 μg/mL  1/50   1/15000 Anti-β60-74 10 μg/mL  1/50   1/2500 Anti-α36-50 10 μg/mL  1/50   1/2500 Anti-α621-635 10 μg/mL  1/50   1/2500 Anti-α171-185biot 10 μg/mL  1/100   1/2500

[0118] Expression of the results of the ELISA tests: for the peptides, the variation in OD between the reactivity of the citrullinated peptide and the reactivity of the non-citrullinated peptide is taken into account, each reactivity being previously subtracted from its blank (OD in the absence of sample). This gives the following formula:

ΔDO=(DO.sub.cit−DO.sub.blank cit)−(DO.sub.noncit−DO.sub.blank noncit)  [Math 1]

[0119] Note that for the fibrinogen, the non-citrullinated form never shows any reactivity and therefore does not need to be evaluated. For the AhFibA test the OD value is therefore simply subtracted from its blank (OD in the absence of sample).

[0120] The ΔDO for the peptides and the OD for the citrullinated fibrinogen are considered to express the titer of the samples tested.

[0121] The inter-assay variations are corrected for by using a range of dilutions of a pool of patient serums tested in each microtiter plate. This reference pool, named P97, was made up of 97 samples of the same volume from 97 patients with a high ACPA titer. The OD values of this internal range are compared to that of reference values where each point in the range is the average of 30 determinations previously performed from this same pool. The least squares method then allows to obtain a correction factor which is applied to all the OD values obtained on the corresponding plate. This corrected OD corresponds to the titer of the sample.

[00001]$\begin{array}{cc}\mathrm{corrected}\mathrm{DO}=\frac{.Math.X*Y*}{.Math.X^2}*\mathrm{DO}& \left[\mathrm{Math}2\right]\end{array}$

[0122] With: X internal values, Y: reference values

[0123] Construction of Chromatography Columns

[0124] Four 1 mL HiTrap Streptavidin HP columns (GE; 29-0513-24) are independently loaded with each of the 4 biotinylated β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides. After equilibration of the column in PBS, a column volume (1 mL) of saturating solution of biotinylated peptides is injected at 0.2 mL/min, incubated for 15 min and then recovered from the column and lyophilised.

[0125] The amount of peptide remaining in the lyophilized fraction is determined by high performance liquid chromatography (UPLC); this amount is subtracted from the starting amount to estimate the amount of peptide bound in the column.

TABLE-US-00003 TABLE 3 Amount of peptide bound Peptide (μmol) β60-74.sub.Cit-NH2 0.83 Ac-α36-50.sub.Cit 1.1 α621-635.sub.Cit 0.87 α171-185.sub.Cit 1.7

[0126] Four columns were loaded with the respective amounts of peptides shown in the table above.

[0127] After use for the serial purification tests, the 4 matrices of the columns loaded with the β60-74.sub.Cit-NH2, α36.sub.Ac-50 .sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides were recovered and mixed and then packed into an empty 10 mL column (GE, C10/10) with volume adapter (GE, AC10), to form a column loaded with the 4 mixed peptides of a final volume of 4 mL.

[0128] Purification

[0129] Two representative pools, P97 (previously described) and P27, a pool of 27 patient serums with an ACPA titer greater than 1.5, as well as individual serums selected on the basis of their reactivity to the 4 β60-74.sub.Cit-NH2, α36.sub.Ac-50 .sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides were diluted to ¼ in PBS before chromatography.

[0130] After equilibration of the columns in PBS at a flow rate of 0.5 mL/min, a volume of 6 mL of pool or serum thus diluted was injected onto the 4 columns mounted in series in the following order: column loaded with β60-74.sub.Cit-NH2 peptide, then α36.sub.Ac-50 .sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit, and subsequently onto the column containing the 4 peptides. Fractions not retained were recovered. After washing in PBS, the elution of the antibodies specific for the 4 peptides was made with a 0.2M pH3 glycine-HCl buffer (Invitrogen; 15527-013). The eluted fractions were recovered and brought to pH 7 with a 2M Tris solution (Euromedex; 77-86-1). Starting serums, non-retained fractions and eluates were then tested in ELISA after correction of the dilution factor induced by chromatography.

[0131] Determination of the Percentage or Rate of Purification

[0132] For each pool or serum treated, the percentage of purification was calculated from the residual immunoreactivity in ACPA (AhFibA ELISA test) of the purified, non-retained fractions and a range of dilutions of the corresponding pool or starting serum.

[00002]$\begin{array}{cc}\%\mathrm{purification}=\left(1-\frac{\mathrm{Dilution}\mathrm{SD}\mathrm{corresponding}\mathrm{to}\mathrm{FNR}\mathrm{OD}}{\mathrm{FNR}\mathrm{experimental}\mathrm{dilution}}\right)*100& \left[\mathrm{Math}3\right]\end{array}$

[0133] A fictitious example of the calculation of percentage purification is shown in FIG. 7 where to obtain an OD equivalent to that of the non-retained fraction (tested at 1/50 after correction for chromatographic dilution) the starting serum must be diluted to 1/400. The concentration of ACPA in the non-retained fraction is therefore 8 times lower than in the starting serum. One eighth, or 12.5%, of the ACPA in the starting serum are present in the non-retained fraction and have therefore not been purified. It can be deduced that seven-eighths of the ACPA in the starting serum, i.e. 87.5%, have been purified: this is the purification rate.

[0134] Reactivity Profile of the Serum Samples

[0135] The reactivity profiles of the 202 samples of patient serums with respect to the citrullinated fibrinogen and the 4 citrullinated peptides were determined by ELISA in at least two independent assays, each being done in duplicate. For the sake of reproducibility, only the microtiter plates whose internal range required the application of a correction factor between 0.5 and 2 with respect to the reference range were considered.

[0136] Results

[0137] ELISA Reactivity Profiles of Serums from Patients with Rheumatoid Arthritis (RA) with Respect to the Citrullinated Fibrinogen (AhFibA Test) and with Respect to the 4 Immunodominant Peptides.

[0138] 202 serums from RA patients were tested in AhFibA and ranked in order of increasing reactivity (OD or titer) (FIG. 1). According to their titer, the patients were divided into quartiles: low (OD=[0.056-0.615]), medium (OD=[0.615-1.161]), high (OD=[1.161-1.876]) and very high (OD≥1.876) titers.

[0139] FIG. 2 shows the percentages of RA serums that are reactive with respect to each of the 4 peptides: 76% were positive for the β60-74Cit-NH2 peptide, confirming that it is the most immunodominant.

[0140] The individual reactivity of the serums with respect to the 4 peptides is highly variable and allows the definition of numerous reactivity profiles. The frequency of the different profiles observed is shown in FIG. 3. Among these 202 serums, 78% are reactive against at least 2 of the 4 peptides, and 95% react with at least one of the 4 peptides. These results show that the use of a combination of the 4 peptides in chromatography should allow to target a very broad spectrum of patients.

[0141] The search for correlations between the reactivities of the serums with respect to the different peptides reveals significant correlations between the β60-.sub.74Cit-NH2, α621-635.sub.Cit and α171-185.sub.Cit peptides; these correlations reflect a certain degree of cross-reactivity between these peptides, but their relative weakness shows that each peptide has its own reactivity. However, no correlation was found between α36.sub.Ac-50.sub.Cit and the other 3 peptides. The double-entry table in FIG. 4 shows the Spearman coefficients calculated for all the peptide pairs compared and whether they are significant or not. These results support the idea that in order to develop a purification system targeting the whole of the ACPA for the largest possible number of patients, it is necessary to use all 4 peptides together. So in the end, this is the choice that is made.

[0142] Purification of Serum Pools from RA Patients on 4 Chromatography Columns Loaded with β60-74.sub.Cit-NH2, .sub.Ac-α36-50 .sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit Peptides Respectively and Mounted in Series, or on a Column Containing a Mixture of the 4 Peptides.

[0143] Two chromatography systems, shown in FIG. 5, were devised to purify the patient serums of the ACPA they contain using the 4 peptides. The first system consists of connecting in series 4 columns loaded respectively with one of the 4 β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit(A) peptides, the second consists of mixing in equivalent quantities the 4 mono-peptide matrices and loading them into a single column (B).

[0144] The efficacy of the two systems was compared using two pools of serums from RA patients. These conditions were chosen because they are more demanding than those for testing an individual serum. Indeed, in a pool containing many serums with different reactivity profiles, the mixture of ACPA is highly polyclonal and composed of ACPA of different antigenic specificity.

[0145] A first pool, P97, consists of 97 serums with a distribution of the AhFibA titres representative of the general population of RA patients. A second pool, P27, consists of 27 serums from patients with high AhFibA titers (OD>1.5).

[0146] FIG. 6 shows the results of chromatography made on these pools, either on the 4 serially mounted columns, loaded with the β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides respectively, or on a column containing the mixture of the 4 peptides. A significant proportion of the ACPA has been absorbed onto the columns and can be eluted from them, while another portion, not absorbed, is found in the fraction not retained by the columns. It appears that both chromatography systems are effective in purifying a large portion of the ACPA contained in the pools and that they give little different results.

[0147] In addition, 10 ACPA-positive serums were selected based on their reactivity profiles with respect to the 4 β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides (Table 4 below).

TABLE-US-00004 TABLE 4 Serum no. β60-74.sub.Cit-NH2 .sub.Ac-α36-50.sub.Cit α621-635.sub.Cit α171-185.sub.Cit n°1 +++ + + n°10 +++ + n°41 +++ ++ n°16 + +++ ++ n°29 + +++ n°25 ++ +++ + n°64 +++ +++ n°22 ++ ++ +++ n°6 +++ +++ + n°17 +++ + +++ +++ (+++: strong positive; ++: medium positive; +: weak positive; empty box: negative)

[0148] These serums as well as the two pools P97 and P27 were run on 4 serial columns loaded with the 4 β60-74.sub.Cit-NH2, .sub.Ac-α36-50.sub.Cit, α621-635.sub.Cit and α171-185.sub.Cit peptides respectively. For all the samples, the ACPA titer decreased sharply after passing through the columns (FIG. 9).

[0149] An efficiency indicator of the purification was needed to compare the results between the different chromatographic methods and samples. As the optical density of a sample in ELISA is not linearly but exponentially correlated with the amount of ACPA contained in this sample, a simple ratio of optical densities between the titer of the starting serum and that of the fraction not retained on the columns could not allow to calculate a purification rate.

[0150] We therefore used a graphical method which consists of determining in the AhFibA ELISA the OD of a series of dilutions of the starting serum, allowing us to establish a reference curve. The OD of the non-retained fraction is determined in the same way, after correction for the chromatography-related dilution factor. The graphical reading on the reference curve indicates to which dilution of the starting serum the OD of the non-retained fraction corresponds. The percentage of purification calculated results from the ratio between the experimental dilution of the non-retained fraction and the corresponding dilution of the starting serum.

[0151] This graphical method of determining the purification rate is illustrated in FIG. 7. An example of the calculation of this rate for one of the treated serums (serum no. 41) is shown in FIG. 8.

[0152] The purification rates calculated for the 2 pools and for the 10 individual serums after a single pass through the columns are all above 65% and 7 out of 12, including the 2 pools, are above 85%. For the serum n° 41, this rate reaches 98%.

[0153] The relationship between the ACPA titer of the starting serum and the purification rate obtained after passage through the 4 columns mounted in series was studied. FIG. 10 shows that there is a strong and significant correlation (r of Spearman: 0.70; p=0.014) between these two parameters. The higher the titer in ACPA of the serum, the more efficient the purification. This counter-intuitive result is due to the fact that the high-titer ACPA have a high affinity and therefore bind more efficiently to their antigens.

REFERENCES

[0154] Throughout this application, various references describe the prior art to which the invention belongs. The disclosures of these references are incorporated by reference into this application.